Regex for non-consecutive upper-case words

Question

Here\'s one for you Regex wizards.

This is for use within Notepad++, i.e. it is entered directly into the search and replace dialog.

I need

Answer 1

You can use this pattern:

(?:[^A-Z]\s+|[^a-zA-Z\s]|\A)\K[A-Z]+\b(?!\s+[A-Z])

Answer 2

I'd use

(?<![[:upper:]]) ([[:upper:]]+) (?!([[:upper:]]| )+)

this will also select anything your locale considers to be uppercase, not just A-Z

Answer 3

I tried that regex and it works for me when I use the 'Mark' feature of the Notepad++ Find/Replace. I'm not sure if you have any other test cases though.

(?:[[:lower:]][^A-Za-z]*|\A)\K\b[[:upper:]]+\b(?=[^A-Za-z]*[[:lower:]]|\Z)

I'm using \A and \Z because you mentioned multilines.

regex101 demo (with some explanations there).

Be sure to check 'Match case'.

Before:

After:

EDIT: As per update, to inverse the highlighting, I came up with that one... a bit lengthy:

[^A-Z]+|(?:(?=[A-Z]*[a-z])\b[A-Za-z]+\b)|(?:\b[A-Z]+\b(?=[^a-z]+[A-Z])(?:[^a-z]*[A-Z]+)*)

regex101 demo

Answer 4

I would try

(^|\.\s+|[a-z]\s+)([A-Z]+)(\.|\s+[a-z]|$)

That matches: a period or start-of-line or end of a lowercase word; followed by an uppercase word; followed by a period, end-of-line or start of a lowercase word.

The word itself is matched in group 2. If Notepad++ supports lookaround assertions, you can do this so that the only captured word is the single capitalized word:

(?:^|\.\s+|[a-z]\s+)([A-Z]+)(?:\.|\s+[a-z]|$)

Answer 5

You should check for surrounding characters with negative lookahead and negative lookbehind:

(?<![A-Z]\s)\b[A-Z]+\b(?!\s[A-Z])

Live demo

Notepad++ v6.5.1 & works on multi-line purposes too